Best printers for Home

Best Mixer Grinder

Best wired Earphones

Best 43 Inch TV in India

Best Wi Fi Routers

Best Home Theatre in India

Smart Watch under 5000

Best Laptops for Education

Best Laptop for Students

<ul class="summary-list"><li>Nvidia's Jensen Huang praised Elon Musk for building a supercomputer cluster in "19 days."</li><li>Huang described the effort as a "superhuman" task, hailing Musk's understanding of engineering.</li></ul><p>Nvidia chief <a target="_blank" class="" href="https://www.businessinsider.com/nvidia-jensen-huang-leadership-style-insiders-tough-boss-admit-mistakes-2024-6">Jensen Huang</a> has praised Elon Musk after his startup xAI built a supercomputer in 19 days, an effort that he said would take others years. <br /><br />In an episode of the "Bg2 Pod" podcast that aired Sunday, Huang said it was a "superhuman" undertaking to build the supercomputer in Memphis in such a short timeframe.</p><p>Huang said, "As far as I know, there's only one person in the world who could do that; Elon is singular in his understanding of engineering and construction and large systems and marshaling resources; it's just unbelievable." </p><p>xAI built its supercomputer, called Colossus, from a cluster of 100,000 Nvidia H100 GPUs. In September, Musk said in an X post that it took 122 days "from start to finish" to bring the training cluster online, appearing to refer to the total project time.<br /><br />The Nvidia CEO also commended xAI's engineering, software, networking, and infrastructure teams, and called them "extraordinary."</p><p>In an interview with Jordan Peterson on X in June, Musk said it took 19 days to get Colossus from hardware installation to beginning training. He added that this was "the fastest by far anyone's been able to do that." <br /><br />Huang also commented on the supercluster: "Just to put it in perspective, 100,000 GPUs—that's easily the fastest supercomputer on the planet as one cluster." <br /><br />He added, "A supercomputer that you would build would take normally three years to plan and then they deliver the equipment and it takes one year to get it all working."</p><p><a target="_blank" class="" href="https://www.businessinsider.com/elon-musk-xai-chips-mark-zuckerberg-2024-9">xAI launched in 2023</a>, and the company released the latest version of its AI chatbot, Grok-2, in August. Earlier in the year, xAI announced that it <a target="_blank" rel="" href="https://www.businessinsider.com/elon-musk-xai-raised-6-billion-funding-vc-saudi-arabia-2024-5"><u>raised $6 billion in a fresh funding round</u></a>.</p><p>Oracle cofounder and chairman Larry Ellison said in September that he had dinner with Musk and Huang at Nobu in Palo Alto, where the longtime friends were <a target="_blank" class="" href="https://www.businessinsider.com/larry-ellison-elon-musk-jensen-huang-gpus-dinner-oracle-nvidia-2024-9">"begging" Huang for GPUs</a>.<br /><br />He said that they told Huang, "Please take our money….we need you to take more of our money," before he quipped, "It went OK, it worked."</p><p>Nvidia and Elon Musk didn't immediately respond to requests for comment from Business Insider, made outside normal working hours.</p>

Jensen Huang hails Elon Musk and xAI for building an AI supercomputer at 'superhuman' speed

<keyword id="399907" type="General" weightage="100" keywordseo="meta" source="Orion">Meta</keyword> (formerly Facebook) has built an <keyword id="9300485" type="General" weightage="20" keywordseo="AI-supercomputer" source="Orion">AI supercomputer</keyword> that, it claims, will be the fastest in the world when it's fully ready in mid-2022.<br /><br />Called the <keyword id="13025237" type="General" weightage="20" keywordseo="AI-Research-SuperCluster" source="Orion">AI Research SuperCluster</keyword> (RSC), the machine is already being used by Meta researchers to train large models in natural language processing (NLP) and computer vision for research, with the aim of training models with trillions of parameters in the near future.<br /><br />"RSC will help Meta's AI researchers build new and better AI models that can learn from trillions of examples; work across hundreds of different languages; seamlessly analyse text, images, and video together; develop new augmented reality tools; and much more," Meta engineers Kevin Lee and Shubho Sengupta said in a statement late on Monday.<br /><br />The first generation of this infrastructure, designed in 2017, has 22,000 NVIDIA V100 Tensor Core GPUs in a single cluster that performs 35,000 training jobs a day.<br /><br />"We wanted this infrastructure to be able to train models with more than a trillion parameters on data sets as large as an exabyte -- which, to provide a sense of scale, is the equivalent of 36,000 years of high-quality video," said Meta researchers.<br /><br />Early benchmarks on RSC, compared with Meta's legacy production and research infrastructure, have shown that it runs computer vision workflows up to 20 times faster and trains large-scale NLP models three times faster.<br /><br />That means a model with tens of billions of parameters can finish training in three weeks, compared with nine weeks before.<br /><br />"RSC is up and running today, but its development is ongoing. Once we complete phase two of building out RSC, we believe it will be the <keyword id="13026718" type="General" weightage="20" keywordseo="fastest-ai-supercomputer" source="Orion">fastest AI supercomputer</keyword> in the world, performing at nearly 5 exaflops of mixed precision compute," said Meta.<br /><br />Through 2022, Meta will work to increase the number of GPUs from 6,080 to 16,000, which will increase AI training performance by more than 2.5x.<br /><br />The storage system will have a target delivery bandwidth of 16 TB/s and exabyte-scale capacity to meet increased demand, the company added.<br /><br /><b>SEE ALSO:</b><br /><br /> <b><a href="https://www.businessinsider.in/cryptocurrency/news/facebook-owner-meta-is-planning-to-allow-users-to-create-and-sell-nfts-amid-booming-market-for-digital-collectibles-report/articleshow/89023785.cms">Facebook-owner Meta is planning to allow users to create and sell NFTs amid booming market for digital collectibles: report</a></b><br /><br /> <b><a href="https://www.businessinsider.in/cryptocurrency/news/google-is-reportedly-building-a-new-blockchain-division-under-a-veteran-search-exec-finally-signaling-interest-in-web3/articleshow/89021923.cms">Google is reportedly building a new blockchain division under a veteran search exec, finally signaling interest in Web3</a></b><br />

Meta claims its AI supercomputer will be the world's fastest

<br /><ul><li>Microsoft has a new <keyword id="6537818" type="UNKNOWN" weightage="20" keywordseo="supercomputer">supercomputer</keyword> that is going help with multi-tasking AI models.</li><li>The company claims it’s the first supercomputer to use <keyword id="561576" type="TECHNOLOGY" weightage="20" keywordseo="cloud-technology">cloud technology</keyword> in order to fine-tune <keyword id="107950" type="TECHNOLOGY" weightage="20" keywordseo="artificial-intelligence">artificial intelligence</keyword> (AI) models.</li><li>It has more than 285,000 CPU cores and 10,000 GPUs, with 400 gigabits per second of network connectivity for each GPU server.</li></ul>Microsoft has a new supercomputer that’s the fifth fastest in the world. However, the company claims that this will be the first supercomputer integrated with the cloud. <br /><br />The computer will be linked to Microsoft’s own Azure infrastructure in order to train large artificial intelligence (AI) models, it announced during the annual <keyword id="9249551" type="UNKNOWN" weightage="20" keywordseo="microsoft-build-2020">Microsoft Build 2020</keyword> conference.<br /><br />The supercomputer is a single system with more than 285,000 CPU cores and 10,000 GPUs. On top of that, each GPU server has 400 gigabits per second of network connectivity.<br /><br />The supercomputer has been built in collaboration and exclusively for the US-based OpenAI, who’s machine learning (ML) efficiency has been getting more efficient by the day — double every 16 months, outpacing even Moore’s law — according to a paper by <a href="https://openai.com/blog/ai-and-efficiency/">Danny Hernandez and Tom Brown</a>, which is yet to peer-reviewed.<br /><br />“We are seeing that larger-scale systems are an important component in training more powerful models,” OpenAI CEO Sam Altman said.<br /><br /><strong>Single task AI models vs multitasking AI models<br /></strong>Historically, data scientists have built smaller AI models that use labelled examples to learn a single task — like translating a language, recognising objects or to deliver the day’s weather report.<br /><br />In a multitasking AI model, one model will be able to meet one than one end goal. For instance, if a model can understand the nuances of language — like the human intent behind a statement — the same algorithm can be used to analyse billions of pages of text, moderate chat content, or even generate code.<br /><br />“This is about being able to do a hundred exciting things in natural language processing at once and a hundred exciting things in computer vision, and when you start to see combinations of these perceptual domains, you’re going to have new applications that are hard to even imagine right now,” said Microsoft’s Chief Technical Officer Kevin Scott, explaining how the application of the supercomputer for AI will the birth of a new class of multitasking AI models.<br /><br /><strong>The recipes for training AI models will soon be open-source<br /></strong>Microsoft’s Turing Model for natural language generation — the largest publicly available AI language model in the world — was released earlier this year. It has over 17 billion parameters. Today, the company announced that it will begin open-sourcing its Turing models along with how-to guides on how to train them using Azure’s Machine Learning.<br /><br />“This has enabled things that were seemingly impossible with smaller models,” said Luis Vargas, a Microsoft partner technical advisor who is spearheading the company’s AI at Scale initiative.<br /><br />This way, not just Microsoft, but other developers will be able to use the models to improve language understanding across its products. <br /><br />To help that long, DeepSeed, an open-source deep-learning library for PyTorch — a scientific computing package — which helps reduce the computing power required to sum large distributed model training, has also gotten an upgrade. According to Microsoft, the model can now train models which are 15 times larger and 10 times faster, as compared to DeepSeed’s earlier version which was released three months ago.<br /><br /><strong>SEE ALSO:<br /><a href="https://www.businessinsider.in/science/health/news/the-worlds-fastest-supercomputer-joins-the-battle-against-coronavirus-to-find-cure/articleshow/74725396.cms">The world's fastest supercomputer joins the battle against coronavirus — and it's speeding up the hunt for a cure</a></strong><br />

Microsoft has a new supercomputer — and it wants to use it to make AI, that can talk like humans do

<p><div class="KonaFilter image-container display-table"><div><div class="image on-image"><p class="source"><span>AMD</span></p><p class="caption">AMD CEO Li</p></div><p></p></div></div></p><ul class="summary-list"><li>AMD is helping Cray Computer build a new supercomputer for the Department of Energy that it promises will be faster than the fastest supercomputers of today. </li><li>The Department of Energy supercomputer, dubbed Frontier, costs more than $600 million and will be operational in 2021.</li><li>The project could help raise AMD's profile in the high-performance computing chip market dominated by Intel and IBM.</li><li><a href="https://www.businessinsider.com/?hprecirc-bullet">Visit Business Insider's homepage for more stories.</a></li></ul><p><span style="font-weight: 400;">AMD is helping Cray Computer build a supercomputer for the Department of Energy that promises to be faster than any other system of its kind in the world. </span></p><p><span style="font-weight: 400;">"If you take the 160 fastest supercomputers today, that's the capability of this machine," Cray CEO Peter Ungaro told Business Insider.</span></p><p><span style="font-weight: 400;">Frontier is expected to boost research in many areas of big data and artificial intelligence, Ungaro s</span></p><p>The supercomputer, dubbed "Frontier" and with a price tag of more than $600 million, is expected to be unveiled Tuesday. It may also give the semiconductor giant a boost in a market dominated by its toughest rivals: high performance chips.</p><p><span style="font-weight: 400;">Frontier promises to be an exascale supercomputer - that is, a system capable of making a quintillion calculations per second - for the Department of Energy's Oak Ridge National Laboratory in Tennessee. Cray CEO Peter Ungaro told Business Insider.</span></p><p><span style="font-weight: 400;">The system will be used for sophisticated simulation and analyses in various fields, including weather, genomics and physics. It is scheduled to start operating in 2021.</span></p><h2><strong>AMD's biggest supercomputer deal</strong></h2><p><span style="font-weight: 400;">AMD CEO Lisa Su said the project is expected to boost the chip giant's competitiveness in important arenas.</span></p><p><span style="font-weight: 400;">"High performance computing and AI are very important for AMD from a strategic standpoint," she told Business Insider. "This is the largest supercomputing deal we've had."</span></p><p><span style="font-weight: 400;">The supercomputer project "significantly improves AMD's standing in the future in high-performance computing and AI," analyst Patrick Moorhead of Moor Insights and Strategy told Business Insider.</span></p><p><span style="font-weight: 400;">"What's interesting is how this is one indicator of what future roadmaps look like, and it appears AMD has a winner," Moorhead, a former AMD corporate vice president, added. </span></p><p><span style="font-weight: 400;">"This specific supercomputer won't be operational until 2021, so this choice wasn't based on what is in the market today, but in 2021. From this we can extract that AMD is competitive," he said. </span></p><h2><strong>Intel dominates this market</strong></h2><p><span style="font-weight: 400;">AMD still lags other chip powerhouses in the high performance space, however.</span></p><p><span style="font-weight: 400;">Intel chips power roughly 95% of the top 500 supercomputers in the world, followed by IBM with around 2%, and AMD and other chip makers after that, according to Gartner analyst Tony Harvey.</span></p><p><span style="font-weight: 400;">"This market space is extremely competitive on performance and cost," he told Business Insider. It is also a "rapidly growing market" that "impacts the data center and cloud markets by being both important for delivering AI and machine learning solutions and proving credibility by demonstrating the performance of the systems in a very large environment."</span></p><p><span style="font-weight: 400;">"AMD being able to demonstrate their performance in this market makes them even more attractive to the cloud vendors, most of whom are already starting to provide AMD-based machines to their clients," Harvey said.</span></p><p class="tagline">Get the latest Intel stock price<a href="http://markets.businessinsider.com/stock/intc-Quote"> here.</a></p>    <div class="margin-top">        <h3>NOW WATCH: <a href="https://www.businessinsider.com/warren-buffett-net-worth-berkshire-hathaway-billionaire-coke-mcdonalds-2019-4">Warren Buffett, the third-richest person in the world, is also one of the most frugal billionaires. Here's how he makes and spends his fortune.</a></h3>        <div>                    </div>    </div>

AMD and Cray just signed a key deal for an incredibly powerful $600 million supercomputer. Here's how it could give AMD an edge on its rival Intel

<ul><li>At the <keyword id="9150718" type="UNKNOWN" weightage="20" keywordseo="build-2020">Build 2020</keyword> developer conference, <keyword id="10387" type="CITY" weightage="20" keywordseo="Microsoft">Microsoft</keyword> unveiled <keyword id="9250000" type="UNKNOWN" weightage="20" keywordseo="Project-Bonsai">Project Bonsai</keyword> and <keyword id="9249999" type="UNKNOWN" weightage="20" keywordseo="Project-Moab">Project Moab</keyword>, its latest products to help engineers develop autonomous industrial control systems.</li><li>With Project Bonsai, engineers can simulate millions of possible methods and implement an optimised process within weeks as opposed to months, without having a background in artificial intelligence (AI).</li><li>For starters in AI, Project Moab allows developers to learn how to teach machines, making it a building block for Project Bonsai.</li></ul>At its Build 2020 developer conference, Windows-maker Microsoft today announced Project Bonsai, a new artificial intelligence (AI) development platform for developing autonomous industrial control systems.<br /><br />This AI platform can be used to make factories more efficient. The use of stuff like robotic arms, forklifts, precision drills and more can be optimised. <br /><br />Microsoft says that Bonsai will be useful across manufacturing, process control, robotics, energy, smart buildings, consumer durable goods and other sectors. Not just that, users can either 3D print the robot or buy one when it goes on sale later this year, Microsoft <a href="https://blogs.microsoft.com/ai-for-business/build-bonsai-public-preview/">said</a> in a statement. <br /><br />There is also a simulation, developed by MathWorks, that developers can try out immediately.<br /><br />  <div type="Article" hostid="261" msid="75831278" title="Microsoft has a new supercomputer — and it wants to use it make AI that can talk like humans do" mstype="2" mssubtype="0"></div><br /><strong>Building autonomous systems without a background in AI</strong><br /><br />The key promise from Project Bonsai is that it allows people without a background in AI to build autonomous systems. This would make it easier for subject matter experts to focus on their subject instead of learning how AI works.<br /><br />Project Bonsai is a result of Microsoft’s acquisition of Bonsai in 2018. It is a “machine teaching” platform ⁠— it uses more than data analytics; it also incorporates real-life experiences and expertise to the solution, as CEO Satya Nadella explained. <br /><br /><div type="image" title="Satya Nadella" msid="75833543" hostid="261" caption="&amp;lt;p&amp;gt;Microsoft CEO Satya Nadella at the Microsoft Future Decoded Tech Summit in Bengaluru&amp;lt;/p&amp;gt;"></div><br />“This approach bridges AI science and software to the traditional engineering world, enabling fields such as chemical and mechanical engineering to build smarter, more capable and more efficient systems by augmenting their own expertise with AI capabilities,” said Mark Hammond, Microsoft general manager for autonomous systems.<br /><br /><strong>Project Moab is a tool for developers to learn how to teach machines</strong><br /><br />If you don’t know how it will fit into your factory or office, you can simulate the physical infrastructure. And if you don’t have a background in AI, there is Project Moab to learn how to. <br /><br /><div type="image" title="Project Moab" msid="75833478" hostid="261" caption="&amp;lt;p&amp;gt;Project Moab&amp;lt;/p&amp;gt;"></div><br />Project Moab is a hardware kit that allows developers to learn how to teach machines to accomplish a certain task. Microsoft is offering it initially as a simulator in MathWorks. Developers will also be able to 3D print robots to learn how this works in different scenarios.<br /><br />Developers and engineers will then be able to use their learnings from Project Moab into building autonomous systems using Project Bonsai.<br /><br />Essentially, Project Moab is the building block of Project Bonsai.<br /><br />  <div type="Youtube" id="6mfUz4PDgOw" title="How to build Intelligent control systems using new tools from Microsoft and simulations by Mathworks" thumbUrl="https://i.ytimg.com/vi/6mfUz4PDgOw/mqdefault.jpg"></div><br /><strong>SEE ALSO:<br /><br /><a href="https://www.businessinsider.in/tech/news/microsoft-has-a-new-supercomputer-and-it-wants-to-use-it-make-ai-that-can-talk-like-humans-do/articleshow/75831278.cms">Microsoft has a new supercomputer — and it wants to use it make AI that can talk like humans do</a><br /><br /></strong><a href="https://www.businessinsider.in/science/health/news/the-worlds-fastest-supercomputer-joins-the-battle-against-coronavirus-to-find-cure/articleshow/74725396.cms"><strong>The world's fastest supercomputer joins the battle against coronavirus — and it's speeding up the hunt for a cure</strong></a><br />

Microsoft says Project Bonsai can make factories more efficient⁠— and you don’t need to know AI, even a mechanical engineer can do it

<p><div class="KonaFilter image-container display-table"><div><div class="image on-image"><p class="source"><span>Ethan Miller / Getty Images</span></p><p class="caption">Nvidia Founder, President and CEO Jensen Huang</p></div><p></p></div></div></p><ul class="summary-list"><li><a href="https://www.businessinsider.com/category/nvidia">Chip maker Nvidia</a> announced the largest acquisition in its history on Monday when it said it planned to buy Mellanox for $6.9 billion cash.</li><li>It's a big win for Nvidia which reportedly jumped into the bidding late and pulled the company away from <a href="https://www.businessinsider.com/category/intel">Intel</a>.</li><li>But the roots of the deal may have actually happened in 2018 when <a href="https://www.businessinsider.com/category/oracle">Oracle</a> sold off its shares in Mellanox.</li></ul><p>Chip maker Nvidia announced the largest acquisition in its history on Monday with its plan to buy Mellanox for $6.9 billion cash.</p><p>Mellanox was such a hot property that Nvidia reportedly won out over other bidders including Intel and Microsoft. Intel's loss was particularly surprising since one of Mellanox's board members, David Perlmutter, was a long-time Intel executive with a 30-year career there. He joined the Mellanox board shortly after he retired from Intel and is now working as a VC for Eucalyptus Growth Capital specializing in Israeli startups. </p><p>Mellanox makes what's known as InfiniBand technology. This is superfast networking technology used for connecting components inside today's biggest, faster "supercomputers," otherwise known as high performance computing (HPC). It is also used in the data centers that support supercomputers.</p><p style="padding-left: 40px;"><em><strong>Read: <a href="https://www.businessinsider.com/former-cisco-execs-sell-startup-to-juniper-for-405-million-2019-3">3 years after launching their first product, these former Cisco execs have sold their startup to Juniper for $405 million</a></strong></em></p><p>This technology was considered to be the next big thing in building faster computers and faster data centers more than a decade ago, possibly even replacing Ethernet as the mainstream standard. But Ethernet got faster (as it always does) and InfiniBand became a niche product for high-performance computing applications. Intel purchased the other major InfiniBand company, QLogic, in 2012 for $125 million.</p><h2>InfiniBand looked like it would simply whither on the vine</h2><p><div class="KonaFilter image-container display-table float_right"><div><div class="image on-image"><p class="source"><span>Paul Sakuma/AP Images</span></p></div><p></p></div></div>Then two things happened to change both companies' futures: big data and its child, artificial intelligence. The rise of big data meant cheaper ways for companies to store and sift through massive amounts of data for insights.</p><p>Artificial intelligence relies on machine learning which relies on big data. ML AI is where computers study massive numbers of examples of a concept in order to learn that concept.</p><p>Nvidia was pulled into the HPC market when its flagship product, graphics cards (originally created to improve video game performance in PCs) were used in these mammoth super computers to improve their performance.</p><p>With the rise of AI, those hefty supercomputer are taking on some of the gnarliest AI problems and Nvidia found a whole new market in HPC. For instance, the fastest computer in the world is an IBM supercomputer <a href="https://www.ibm.com/thought-leadership/summit-supercomputer/">known as the Summit.</a> It models supernovas, new materials, cures for cancer, genetics and the environment.</p><p>And this one supercomputer uses a cool 27,648 of NVIDIA's GV100 graphics cards as well as 4,600 Mellanox network interface cards, <a href="https://s22.q4cdn.com/364334381/files/doc_presentations/2019/03/NVIDIA-to-Acquire-Mellanox.pdf">Nvidia says.</a></p><p style="padding-left: 40px;"><em><strong>Read: <a href="https://www.businessinsider.com/obama-shares-tricks-for-cool-under-pressure-2019-3">'No drama Obama' shared his tricks for staying cool under pressure</a></strong></em></p><p>Nvidia dove headfirst into the AI chip market with an array of specialized products and is now considered the market leader. It has worked closely with Mellanox for years. Nvidia's cards and chips are used in 127 of the Top 500 fastest supercomputers in the world while Mellanox's network tech is used in 265 of them, Nvidia says.</p><p>Meanwhile, at CES in January Intel ramped up the competitive pressure on Nvidia by launching its own AI chips to challenge Nvidia's lead. Intel's Nervana Neural Network Processor for Inference (NNP-I) product was created with help from Facebook is slated for release in the second half of 2019, reports<a href="https://www.datacenterknowledge.com/machine-learning/intel-steps-its-challenge-nvidia-s-ai-chip-dominance-facebook-s-help"> the Datacenter Knowledge trade news site.</a></p><h2>A happy ending for, almost, everyone</h2><p>When Nvidia got wind that Intel was in talks to buy Mellanox, it essentially had no choice but to launch a competing offer: an Intel-Mellanox combination would have given Intel a stranglehold on the InfiniBand market. So Nvidia jumped into the bidding, reported <a href="https://www.calcalistech.com/ctech/articles/0,7340,L-3758001,00.html">Israeli newspaper CTech.</a></p><p>Nvidia says that the combined companies are looking at a $60 billion market. Had Nvidia lost the bid, Intel, its biggest rival in AI chips could have walked with Mellanox and put a near stranglehold on that $60 billion market as well.</p><p>Interestingly, this whole sweet deal for Mellanox shareholders had its roots in pressure from an activist investor in early 2018. That's when <span>Starboard took an interest in Mellanox, tried to oust all of its directors and by June, managed to get <a href="https://www.reuters.com/article/us-mellanox-settlement-starboard/mellanox-starboard-settle-on-new-board-members-idUSKBN1JF2Z6">three of its own board members </a>appointed.</span></p><p>As this board fight took place, Oracle sold at least half of its stake taken in Mellanox. Oracle bought the 10% stake in Mellanox in 2010 (the days when InfiniBand was thought to be the next big networking thing) at around $20 a share In 2018, Oracle sold its stake for $63.90. By selling those shares, Oracle gave leverage to the activist investors who wanted to sell the company, <a href="https://www.thestreet.com/story/14461577/1/oracle-selling-mellanox-stake-helps-starboard.html">Ronald Orol at The Street reported.</a></p><p>Fast forward to March, 2019 and Mellanox has officially gone to Nvidia's open arms for $125 per share in cash. It's a happy ending for all, except, perhaps, Intel.</p><p class="tagline">Get the latest Intel stock price<a href="http://markets.businessinsider.com/stock/intc-Quote"> here.</a></p>    <div class="margin-top">        <h3>NOW WATCH: <a href="https://www.businessinsider.com/samsung-foldable-phone-galaxy-x-f-winner-2018-9">Everything we know about Samsung's foldable phone</a></h3>        <div>                    </div>    </div>

How Oracle inadvertently helped Nvidia spend $6.9 billion to win a deal away from Intel

<p><div class="KonaFilter image-container display-table"><div><div class="image on-image"><p class="source"><span>AP Photo/Michael Sohn</span></p><p class="caption">Sundar Pichai, CEO of Google, speaks during a statement as part of the opening of a new Google office in Berlin, Germany, Tuesday, Jan. 22, 2019.</p></div><p></p></div></div></p><ul class="summary-list"><li>Google scientists have claimed a major breakthrough in superfast computing after running an experiment which appears to demonstrate just how much faster quantum computers would be compared to classical ones.</li><li>Quantum computing is complicated, but based on the idea that computers can incorporate aspects of quantum theory. In theory, computational tasks can be carried out exponentially faster by quantum computers than by classical computers. This is known as quantum supremacy.</li><li>Google's scientists tested a 54-qubit (quantum bit) processor named Sycamore, and it performed a target computation in 200 seconds. To put this into perspective, the scientists estimate it would take the world's fastest supercomputer 10,000 years to produce a similar output.</li><li>IBM, which is also working on quantum computing research, cast doubt on Google's findings and said on Monday that quantum supremacy "has not been met."</li><li><a href="https://www.businessinsider.com/international?r=US&amp;IR=T">Visit Business Insider's homepage for more stories.</a></li></ul><p>Google scientists say they've made a breakthrough in one of the most cutting-edge areas of computing, publishing a paper that demonstrates how much faster quantum computers could be versus traditional computers.</p><p>According to a paper <a href="https://www.nature.com/articles/s41586-019-1666-5">published in Nature</a>, an experimental quantum processor developed by the scientists took seconds to carry out computation which they claim the world's fastest supercomputer would need 10,000 years to carry out. This is known as quantum supremacy, and would mark a major breakthrough in the experimental field.</p><p>Quantum computing is complex and very much at the research stage. It involves computing technology that is based on the principles of quantum theory. </p><p>Instead of bits, which today's computers use to process information, quantum computers use quantum bits, or 'qubits.' Where classical computers are constrained by how fast they can switch between outputting zeros and outputting ones, quantum computers would face no such constraints, because qubits can theoretically exist in both states simultaneously.</p><p>The upshot is that quantum computers would be able to carry out particular tasks exponentially faster than a standard computer - millions of times faster. Google says that this is what it's demonstrated.</p><p>John Martinis and Sergio Boixo, two senior scientists at Google AI Quantum, <a href="https://ai.googleblog.com/2019/10/quantum-supremacy-using-programmable.html">wrote on Wednesday</a> that they were motivated to run the experiment to find out if quantum computing "will ever do something useful" and whether it's "worth investing in."</p><p>Elaborating on their motivations, they wrote: "For such large-scale endeavors, it is good engineering practice to formulate decisive short-term goals that demonstrate whether the designs are going in the right direction. So, we devised an experiment as an important milestone to help answer these questions."</p><p>The scientists' experiment involved building a 54-qubit (quantum bit) processor named Sycamore, made up of fast, high-fidelity quantum logic gates.</p><p><em><strong>Read more:</strong> <a href="https://www.businessinsider.com/quantum-computing-explained-2019-5?r=US&amp;IR=T">Quantum computing could change everything, and IBM is racing with Microsoft, Intel, and Google to conquer it. Here's what you need to know.</a></em></p><p>They then tested Sycamore's computing ability, testing how long it would take to perform a single target calculation. Setting the bar for success high, the scientists agreed on a "sensitive computational benchmark that fails if just a single component of the computer is not good enough."</p><p>According to the scientists, Sycamore performed the target calculation in 200 seconds. They claimed that, according to their measurements, it would take the world's fastest supercomputer 10,000 years to produce a similar output.</p><p><div class="KonaFilter image-container display-table"><div><div class="image on-image"><p class="source"><span>Flickr/IBM</span></p><p class="caption">IBM disputes Google's claims (pictured is an IBM quantum computing researcher).</p></div><p></p></div></div></p><p>One expert says that Google's claims are "compelling."</p><p>Peter Knight, a senior physicist at Imperial College London, <a href="https://www.bbc.co.uk/sounds/play/m0009jhk">told the BBC's Today Programme</a> on Wednesday that Google's experiment was "a real breakthrough," and agreed that this was "the real deal" in quantum computing research.</p><p>Though Knight noted similar research had been conducted by IBM, he "always felt it was likely the [Google] group would be there first." A draft paper on the same Google experiment <a href="https://www.ft.com/content/b9bb4e54-dbc1-11e9-8f9b-77216ebe1f17">reportedly appeared</a> on the NASA website in August before being swiftly taken down. </p><p> As quantum computing remains very much in its infancy, it's difficult to imagine how and in what ways it might impact human society, though no-one disputes <a href="https://www.businessinsider.com/quantum-computers-will-change-the-world-2015-4?r=US&amp;IR=T">its potential to impact just about everything</a>. Knight briefly touched on one area it may affect: data security. Though he said the potential threat posed by quantum computing to processes such as encryption is still "distant," the threat "is still there."</p><p>For its part, IBM disputes Google's putative breakthrough. <a href="https://www.ibm.com/blogs/research/2019/10/on-quantum-supremacy/">In a blog post on Monday</a>, they claimed "the goal [of quantum supremacy] has not been met."</p> <p class="tagline">Get the latest Google stock price<a href="http://markets.businessinsider.com/stock/GOOG-Quote"> here.</a></p>    <div class="margin-top">        <h3>NOW WATCH: <a href="https://www.businessinsider.com/watch-google-pixel-4-event-in-12-minutes-2019-10">Watch Google's Pixel 4 event in 12 minutes</a></h3>        <div>                    </div>    </div>

Google scientists have claimed a massive breakthrough in cutting-edge computing with 'quantum supremacy'

<p class="article-excerpt"><div class="KonaFilter image-container display-table float_right"><div><div class="image on-image"></div><p></p></div></div>An array of science-fiction-like technologies would likely make their debut</p><p>U.S. and Chinese warships battle at sea, firing everything from cannons to cruise missiles to lasers. Stealthy Russian and American fighter jets dogfight in the air, with robotic drones flying as their wingmen. Hackers in Shanghai and Silicon Valley duel in digital playgrounds.</p><p>And fights in outer space decide who wins below on Earth.</p><p>Are theses scenes from a novel or what could actually take place in the real world the day after tomorrow? The answer is both.</p><p>Great power conflicts defined the 20th century: Two world wars claimed tens of millions of lives, and the Cold War that followed shaped everything from geopolitics to sports. But at the start of the 21st century, the ever-present fear of World War III seemed to be in our historic rearview mirror.</p><p>Yet that risk of the past has made a dark comeback. Russian land grabs in Ukraine and constant flights of bombers decorated with red stars probing Europe's borders have put NATO at its highest levels of alert since the mid 1980s. In the Pacific, the U.S. and a newly powerful and assertive China are engaged in a massive arms race. China built more warships and warplanes than any other nation during the last several years, while the</p><p>Pentagon just announced a strategy to "offset" it with a new generation of high-tech weapons. Indeed, it's likely China's alleged recent <a href="http://www.nbcnews.com/tech/security/clapper-china-leading-suspect-opm-hack-n381881">hack</a> of federal records at the Office of Personnel Management was not about cyber crime, but a classic case of what is known as "preparing the battlefield," gaining access to government databases and personal records just in case.</p><p>The worry is that the brewing 21<span>st</span> century Cold War with China and its junior partner Russia could at some point turn hot. "A U.S.-China war is inevitable" recently warned the <a href="http://www.japantimes.co.jp/news/2015/05/25/asia-pacific/china-state-media-warns-war-south-china-sea-unless-u-s-backs/#.VWXDM-uJndl">Communist Party's official People's Daily newspaper</a> after recent military face-offs over rights of passage and artificial islands built in disputed territory.</p><p>This may be a bit of posturing both for U.S. policymakers and a highly nationalist domestic audience: A 2014 poll by the Perth U.S.-Asia center found that <a href="http://perthusasia.edu.au/usac/assets/media/docs/publications/2014_Exploring_Chinas_Maritime_Consciousness_Final.pdf">74% of Chinese think their military would win in a war with the U.S</a>. But it points to how the global context is changing. Many Chinese officers have begun to lament out loud what they call "peace disease," their term for never having served in combat.</p><p><div class="KonaFilter image-container display-table"><div><div class="image on-image"></div><p></p></div></div></p><p>Wars start through any number of pathways: One world war happened through deliberate action, the other was a crisis that spun out of control. In the coming decades, a war might ignite accidentally, such as by two opposing warships trading paint near a reef not even marked on a nautical chart. Or it could slow burn and erupt as a reordering of the global system in the late 2020s, the period at which China's military build up is on pace to match the U.S.</p><p>Making either scenario more of a risk is that military planners and political leaders on all sides assume their side would be the one to win in a <a href="http://news.usni.org/2014/02/18/navy-official-china-training-short-sharp-war-japan">"short" and "sharp"</a> fight, to use common phrases. It would be anything but.</p><p>A great power conflict would be quite different from the small wars of today that the U.S. has grow accustomed to and, in turn, others think reveal a new American weakness. Unlike the Taliban or even Saddam's Iraq, great powers can fight across all the domains; the last time the U.S. fought a peer in the air or at sea was in 1945. But a 21st century fight would also see battles for control of two new domains.</p><p>The lifeblood of military communications and control now runs through space, meaning we'd see humankind's first battles for the heavens. Similarly, we'd learn "cyber war" is far more than stealing Social Security Numbers or e-mail from gossipy Hollywood executives, but the takedown of the modern military nervous system and <a href="http://www.vanityfair.com/news/2011/04/stuxnet-201104">Stuxnet</a>-style digital weapons. Worrisome for the U.S. is that last year, the Pentagon's weapons tester found <a href="http://in.reuters.com/article/2015/01/21/us-cybersecurity-pentagon-idINKBN0KU02920150121?feedType=RSS&amp;feedName=everything&amp;virtualBrandChannel=11709">nearly every single major weapons</a> program had "significant vulnerabilities" to cyber attack.</p><p><div class="KonaFilter image-container display-table"><div><div class="image on-image"></div><p></p></div></div></p><p>A total mindshift is required for this new reality. In every fight since 1945, U.S. forces have been a generation ahead in technology, having uniquely capable weapons like nuclear-powered aircraft carriers. It has not always translated to decisive victories, but it has been an edge every other nation wants.</p><p>Yet U.S. forces can't count on that "overmatch" in the future. These platforms are not just vulnerable to new classes of weapons like long-range missiles, but <a href="http://mobile.businessweek.com/articles/2014-11-26/as-the-u-dot-s-dot-cuts-r-and-d-spending-china-is-raising-its-stake">China, for example, overtook the EU in R&amp;D spending last year and is on pace to match the U.S. within five years</a>, with new projects ranging from the world's fastest supercomputers to three different long-range drone-strike programs.</p><p>And now off-the-shelf technologies can be bought to rival even the most advanced tools in the U.S. arsenal. The winner of a recent robotics test, for instance, was not a U.S. <keyword id="6354810" type="ORGANIZATION" weightage="100" keywordseo="Defense">defense</keyword> contractor but a group of <a href="http://www.fastcompany.com/3047155/south-korean-team-takes-2-million-top-prize-at-darpa-robotics-challenge">South Korea student engineers</a>.</p><p>An array of science-fiction-like technologies would likely make their debut in such a war, from AI battle management systems to autonomous robotics. But unlike the ISIS's of the world, great powers can also go after high-tech's new vulnerabilities, such as by hacking systems and knocking down GPS.</p><p><div class="KonaFilter image-container display-table"><div><div class="image on-image"></div><p></p></div></div></p><p>The recent steps taken by the U.S. Naval Academy illustrate where things might be headed. It added a cybersecurity major to develop a new corps of digital warriors, and also requires all midshipmen learn celestial navigation, for when the high tech inevitably runs into the age old fog and friction of war.</p><p>While many leaders on both sides think any clash might be geographically contained to the straights of Taiwan or the edge of the Baltic, these technological and tactical shifts mean such a conflict is more likely to reach into each side's homelands in new ways. Just as the Internet reshaped our notions of borders, so too would a war waged partly online.</p><p>The civilian players would also be different than those in 1941. The hub of any war economy wouldn't be Detroit. Instead, tech geeks in Silicon Valley and shareholders in Bentonville, Ark., would wrestle with everything from microchip shortages to how to retool the logistics and allegiance of a multinational company. The new forms of civilian conflict actors like Blackwater private military firms or Anonymous hacktivist groups are unlikely to just sit out the fight.</p><p><div class="KonaFilter image-container display-table"><div><div class="image on-image"></div><p></p></div></div></p><p>A Chinese officer argued in a regime paper, <a href="http://en.people.cn/n/2014/0916/c90000-8782955.html">"We must bear a third world war in mind when developing military forces."</a> But there is a far different attitude in Washington's defense circles. As the U.S. Chief of Naval Operations worried last year, "If you talk about it openly, you cross the line and unnecessarily antagonize. You probably have a sense about how much we trade with that country, it's astounding."</p><p class="p1">This is true, but both the historic trading patterns between great powers before each of the last world wars and the risky actions and heated rhetoric out of Moscow and Beijing over the last year demonstrate it is no longer useful to avoid talking about the great power rivalries of the 21st century and the dangers of them getting out of control.</p><p class="p1">We need to acknowledge the real trends in motion and the real risks that loom, so that we can take mutual steps to avoid the mistakes that could create such an epic fail of deterrence and diplomacy. That way we can keep the next world war where it belongs, in the realm of fiction.</p><p>(Written by P.W. Singer and August Cole)</p>    <div class="margin-top">        <h3>NOW WATCH: <a href="http://www.businessinsider.com/new-omega-globemaster-watch-unboxing-2015-5">Forget the Apple Watch - here's the new watch everyone on Wall Street wants</a></h3>        <div>            <div>        <div class="ooyala-video-player" data-video-autoplay="false" data-video-id="5iaDNwdDrYYMm2yipFb14ndbRkPnemqO" data-player-variable="ooyalapopvideo" id="ooyalaplayer_popvideo" style="width:1280px;height:720px"></div>        </div>        </div>            </div>

You Searched For "fastest ai supercomputer" and got 8 results

Jensen Huang hails Elon Musk and xAI for building an AI supercomputer at 'superhuman' speed

Jyoti Mann

Meta claims its AI supercomputer will be the world's fastest

IANS

Microsoft has a new supercomputer — and it wants to use it to make AI, that can talk like humans do

Prabhjote Gill

AMD and Cray just signed a key deal for an incredibly powerful $600 million supercomputer. Here's how it could give AMD an edge on its rival Intel

Benjamin Pimentel

Microsoft says Project Bonsai can make factories more efficient⁠— and you don’t need to know AI, even a mechanical engineer can do it

Rounak Jain

How Oracle inadvertently helped Nvidia spend $6.9 billion to win a deal away from Intel

Julie Bort

Google scientists have claimed a massive breakthrough in cutting-edge computing with 'quantum supremacy'

Charlie Wood

What World War III would probably look like

time

Popular Keywords

Buying Guides